异构Redis集群大规模评论数据存储负载均衡设计

发布时间：2018-01-21 01:49

本文关键词： 大规模评论数据存储负载均衡查询优化　出处：《华东师范大学学报(自然科学版)》2017年05期 　论文类型：期刊论文

【摘要】：大规模评论数据的存储与查询性能对构建于其上的各类应用的快速响应具有重要影响.同时,异构计算环境中各计算节点性能呈现差异,如何充分开采各节点的计算和存储性能,优化大规模评论数据的存储与查询性能,是一个关键挑战.基于Redis集群的数据管理优势,首先提出了一种同构环境下基于卡槽存储平衡的大规模评论数据存储模型;然后论证了卡槽数目与节点查询效率的关系,以"负载与访问性能相平衡"的原则分配卡槽,进一步设计了异构环境下的集群节点负载计算和存储分配方法,充分开采了异构Redis集群中不同节点的性能.实验结果表明,提出的存储模型具有很好的存储平衡效果,提升了集群的整体查询效率.
[Abstract]:The storage and query performance of large-scale comment data has an important impact on the rapid response of various applications built on it. At the same time, the performance of computing nodes in heterogeneous computing environments is different. How to fully exploit the computing and storage performance of each node and optimize the storage and query performance of large-scale comment data is a key challenge. Firstly, a large scale review data storage model based on slot storage balance in isomorphic environment is proposed. Then, the relationship between the number of card slots and the efficiency of node query is demonstrated. According to the principle of "balance between load and access performance", the load calculation and storage allocation method of cluster nodes in heterogeneous environment is further designed. The performance of different nodes in heterogeneous Redis cluster is fully exploited. Experimental results show that the proposed storage model has a good storage balance effect and improves the overall query efficiency of the cluster.
【作者单位】：桂林电子科技大学广西可信软件重点实验室;桂林电子科技大学广西自动检测技术与仪器重点实验室;
【基金】：国家自然科学基金(61363005,61462017,U1501252) 广西自然科学基金(2014GXNSFA A118353,2014GXNSFAA118390) 广西自动检测技术与仪器重点实验室基金(YQ15110) 广西高校中青年教师基础能力提升项目(ky2016YB156)
【分类号】：TP333
【正文快照】： o引言用户在电子商务网站、Web论坛等贡献了海量的评论数据,其蕴含了用户的兴趣偏好、主观观点、潮流取向等信息,对在线推荐、舆情监控、用户画像、公司商品服务决策等分析应用具有重要价值.上述分析应用通常要求以用户、评论对象、时间段等维度为查询约束开展高10的数据获取

【相似文献】